[liberationtech] The Invention of "Ethical AI"

Thu Dec 26 22:13:27 CET 2019

Anonymity doesn’t protect “Particular Social Groups” and aggregate anonymous data analysis and mining is the basis for discrimination of entire segments of the population! Like zip code discrimination. Food deserts. Etc.

Regards / Saludos / Grato

Andrés Leopoldo Pacheco Sanfuentes

> On Dec 26, 2019, at 2:01 PM, L Jean Camp <ljeanc at gmail.com> wrote:
> 
> 
> John, your "point" is ahistorical to the point of being silly. Disinterest does not justify derailing. 
> 
> Steve, 
> In terms of the HUD rule. It is an actual rule-making addressing the role of AI and the consequent responsibility for harm. The anonymity proposal holds real promise but .... would it work?  Could the use anonymous data would help here because of the deep, profound, and thus highly correlated and consistently significant role, reality of race and wealth in the US. 
> 
> The HUD proposal holds any entity that causes harm to be protected from any risk-based or harm-based remedies regardless of the magnitude of the harm. That is highly unusual in any domain. Here is a pointer to a description of how risk-based versus harm-based regulations have applied in security and privacy. (https://papers.ssrn.com/sol3/papers.cfm?abstract_id=2242474 low math version http://www.ljean.com/files/eCrime2015CameraReady.pdf). 
> 
> I think the same approach could be used in considering ethical AI -- are there areas where the risk are well understood? I do not know of any. So making entities judgement-proof for use of anonymous data when we do not understand the implications is not something I would agree with.
> 
> I am thinking of another area where anonymity would not help. For example, in the use of sensing and responding for collision avoidance the CAMP standard (coincidence of course) implements a PKI where the keys rapidly rotate for privacy. Does this create privacy? I think not, I think a . better protocol is superior to rotating keys. Does this create an ability of any vehicle under the system to use legitimate keys to implement a traffic-stopping sybil attack without breaking any crypto, just code abuse? That is very implementation specific. And AI is deeply embedded into the decisions of the system. I picked this example because automobiles is an area that is highly regulated, and the physical risks are very well understood. But add crypto and AI, then it can break in unexpected ways.
> 
> I would love to hear your thoughts on this second situation.  Thank you. 
> 
>> On Thu, Dec 26, 2019 at 1:05 PM John Young <jya at pipeline.com> wrote:
>> "Ethical" is a marketing, manipulative term, applied to exploitive, 
>> deceptive initiatives. While a few who adopt it may have beneficial 
>> intentions, best to be wary of evolution of initial good will 
>> transformed into clothing PR-extolled shenanigans.
>> 
>> There is a long history of ethics, philosophically and practically, 
>> the purpose its invention, used to cosmetize and flummox, heirarchy 
>> simulating equality. If a venture needs ethical glossing, it is 
>> confessing aim to allure and entrap. In that sense, ethical AI, is 
>> bogus, faith-based.
>> 
>> The critique below is properly skeptical.
>> 
>> At 04:33 AM 12/26/2019, you wrote:
>> >On Mon, Dec 23, 2019 at 06:06:14PM -0800, Steve Phillips wrote: > 
>> >But "AI ethics" is a much more vague notion, and not one I've seen > 
>> >connected to meaningful action. That seems to be the point, to take 
>> >it on a meta level such that there is no effective action. Yet I see 
>> >a pretty simple and straightforward way to make AI ethical: AI must 
>> >*never* be fed private, personal or otherwise non-anonymized data of 
>> >human beings, not even under the auspices of "consent". Only then 
>> >can AI do the useful things like predicting the spread of diseases 
>> >without turning human society into a chess game for the owners of 
>> >the AI. Those who are allowed to grab the data become the masters of 
>> >the planet. Even if you kid yourself that you own your own data. 
>> >This principle of uncollectibility of personal data needs to be cast 
>> >into regulation, refined and defined in a sane way as it is of 
>> >higher importance than democratic constitutions. -- Liberationtech 
>> >is public & archives are searchable from any major commercial search 
>> >engine. Violations of list guidelines will get you moderated: 
>> >https://lists.ghserv.net/mailman/listinfo/lt. Unsubscribe, change to 
>> >digest mode, or change password by emailing lt-owner at lists.liberationtech.org.
>> 
>> 
>> 
>> -- 
>> Liberationtech is public & archives are searchable from any major commercial search engine. Violations of list guidelines will get you moderated: https://lists.ghserv.net/mailman/listinfo/lt. Unsubscribe, change to digest mode, or change password by emailing lt-owner at lists.liberationtech.org.
> 
> 
> -- 
> Prof. L. Jean Camp
> http://www.ljean.com
> 
> 
> Research Gate: https://www.researchgate.net/profile/L_Camp
> DBLP: http://dblp.uni-trier.de/pers/hd/c/Camp:L=_Jean
> SSRN: https://papers.ssrn.com/sol3/cf_dev/AbsByAuth.cfm?per_id=262477
> Scholar: https://scholar.google.com/citations?user=wJPGa2IAAAAJ
> Make a Difference 
> http://www.ieeeusa.org/policy/govfel/congfel.asp
> -- 
> Liberationtech is public & archives are searchable from any major commercial search engine. Violations of list guidelines will get you moderated: https://lists.ghserv.net/mailman/listinfo/lt. Unsubscribe, change to digest mode, or change password by emailing lt-owner at lists.liberationtech.org.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.ghserv.net/pipermail/lt/attachments/20191226/2da70a9e/attachment-0001.html>